U.S. government restricts AI cyber model access to CISA-approved entities

The U.S. government will implement a formal policy by Q1 2027 that restricts access to advanced AI cybersecurity models like GPT-5.5-Cyber and Mythos to entities approved by the Cybersecurity and Infrastructure Security Agency (CISA). This will include mandatory background checks and usage audits.

Implied probability (Yes)

55%